1. The Need for a New Radio: Why LTE's Air Interface Wasn't Enough

The 4G LTE air interface was a masterpiece of its time. Based on OFDMA and SC-FDMA, it provided a robust and efficient platform that successfully delivered the mobile broadband revolution. However, as the vision for the next generation began to take shape, it became clear that the rigid structure of the LTE radio, as well-designed as it was, could not accommodate the vastly diverse and often conflicting demands of the future. The very definition of "mobile communication" was expanding beyond just faster smartphones.

The limitations of the LTE air interface were tied to its "one-size-fits-all" design. Key parameters of the radio signal were fixed, optimized primarily for the mobile broadband use case. For instance, the subcarrier spacing was always $15$ kHz, and the frame and slot structures were largely inflexible. This design was excellent for delivering fast data to phones, but it was not a good fit for the emerging use cases that would define 5G:

• Ultra-Reliable Low-Latency Communications (URLLC): Mission-critical applications like autonomous vehicles or remote surgery require latencies of just a single millisecond. The fixed timing structure of LTE simply could not support such near-instantaneous communication.
• Massive Machine-Type Communications (mMTC): The IoT demands connectivity for billions of simple, low-power sensors. The LTE radio was too complex and power-hungry for a device that might only need to send a few bytes of data per week and last for a decade on a battery.
• Diverse Spectrum Bands: The future of mobile broadband (eMBB) required harnessing vast new swaths of spectrum in the high-frequency millimeter wave (mmWave) bands. The fixed LTE radio parameters, optimized for sub-6 GHz frequencies, were not well-suited for the unique physical properties of mmWave signals.

It became clear that a new, fundamentally more flexible and adaptable air interface was needed. It couldn't just be an evolution; it had to be a ground-up redesign capable of dynamically molding itself to fit any service and any frequency band. This is the genesis of 5G New Radio (NR).

2. Core Philosophy of 5G NR: Native Flexibility and Forward Compatibility

5G NR was designed with a core philosophy that sets it apart from all previous cellular technologies: it is inherently flexible and designed to be future-proof. Rather than creating a rigid standard, the designers at 3GPP created a highly configurable toolkit of radio parameters and waveforms that can be mixed and matched to meet the specific requirements of any application or deployment scenario.

Key aspects of this philosophy include:

• Scalability Across All Frequencies: 5G NR is designed to operate seamlessly across an enormous range of frequencies, from low bands below 1 GHz (for wide area coverage), through the traditional mid-bands (1-6 GHz), all the way up to the high-frequency mmWave bands (above 24 GHz).
• Flexible Waveform Numerology: This is the heart of 5G NR's adaptability. Instead of a fixed radio waveform, NR introduces a flexible set of parameters (the "numerology") that can be dynamically changed to optimize the signal for latency, reliability, or spectral efficiency.
• Unified Air Interface for All Services: 5G NR provides a single, unified radio framework that can efficiently support all three core 5G use cases (eMBB, URLLC, and mMTC).
• Forward Compatibility: The standard was designed with future evolution in mind. Its structure allows for new features, new frequency bands, and new service types to be added in future releases without breaking compatibility with existing devices and networks.

3. The Cornerstone of Flexibility: 5G NR Numerology

The single most important and innovative concept in the 5G NR air interface is its flexible numerology. In the context of a radio waveform, numerology refers to the fundamental parameters that define its structure in time and frequency. While LTE was built on a single, fixed numerology, 5G NR defines a family of scalable numerologies.

What is Numerology?

The 5G NR waveform, like LTE's, is based on OFDM (Orthogonal Frequency-Division Multiplexing). A numerology in this context is defined by a specific Subcarrier Spacing (SCS). Changing the SCS has a direct, inverse relationship on the duration of other key waveform parameters:

• Symbol Duration: The period of time used to transmit a single OFDM symbol. If you double the subcarrier spacing, the symbol duration is halved.
• Cyclic Prefix (CP) Duration: A small copy of the end of an OFDM symbol that is prepended to the beginning. The CP acts as a guard interval to prevent inter-symbol interference caused by multipath delays. Its duration also scales inversely with the SCS.

The Scalable Numerologies of 5G NR

5G NR defines a set of numerologies based on scaling the foundational 15 kHz subcarrier spacing of LTE by powers of two. The standard specifies a parameter, $\mu$ (mu), to define the SCS:

This results in the following primary numerologies:

4. The Trade-offs: Matching Numerology to the Service

The choice of which numerology to use is not arbitrary; it involves fundamental trade-offs between latency, efficiency, and robustness. The ability to select the right numerology for the right job is the genius of 5G NR.

5. The 5G NR Frame Structure: Time and Frequency Resources

The 5G NR organizes its radio resources into a flexible grid of time and frequency, providing a structure for scheduling data.

The Time Domain: Frames, Subframes, and Slots

The basic timing structure is built around a $10$ ms radio frame, just like in LTE. However, how that frame is subdivided is much more flexible.

• Radio Frame: The largest time unit, fixed at $10$ ms long.
• Subframe: Each frame is divided into 10 subframes of $1$ ms each.
• Slot: Here is where the flexibility of numerology comes in. A subframe is composed of one or more slots. The number of slots per subframe depends on the subcarrier spacing used:
  • With 15 kHz SCS ($\mu=0$), there is 1 slot per subframe.
  • With 30 kHz SCS ($\mu=1$), there are 2 slots per subframe.
  • With 60 kHz SCS ($\mu=2$), there are 4 slots per subframe.
  • With 120 kHz SCS ($\mu=3$), there are 8 slots per subframe.
• OFDM Symbols: In the most common configuration, each slot consists of 14 OFDM symbols. The actual time duration of these 14 symbols becomes shorter as the SCS increases.
• Mini-Slots: To support even lower latencies for URLLC, 5G NR introduces the concept of a mini-slot. This allows the network to schedule a transmission for as few as 2, 4, or 7 OFDM symbols, without having to wait for a full 14-symbol slot to become available.

6. The Frequency Domain: Resource Blocks and Bandwidth Parts (BWP)

The frequency dimension of the radio grid is also highly adaptable.

• Resource Block (RB): The smallest unit of frequency resources that can be allocated to a user is a Resource Block. An RB consists of 12 contiguous subcarriers. For a given numerology, all 12 subcarriers in an RB have the same spacing. The scheduler in the gNB allocates data to users in units of RBs and slots.
• Bandwidth Part (BWP): This is another key innovation of 5G NR. In LTE, a device had to be capable of receiving the entire channel bandwidth (e.g., 20 MHz) at all times, even if it was only receiving a small amount of data. This was inefficient from a power consumption perspective. A BWP is a subset of the total channel's Resource Blocks. The network can configure one or more BWPs for a device. The device only needs to actively monitor and process the bandwidth of its currently active BWP, not the full carrier bandwidth. For instance, in a $100$ MHz wide carrier, a device in an idle or low-activity state might be configured with a narrow $20$ MHz BWP. When a high-speed data transfer is needed, the network can instantly instruct the device to switch to the full $100$ MHz BWP. This allows for significant battery savings and also enables low-cost, simpler devices to operate on a wideband 5G carrier without needing the expensive RF hardware to support the full bandwidth.